Hybrid circular-square complementary split-ring resonator sensor for monitoring abnormal blood protein levels for disease diagnosis.

The proposed work aims to develop a compact, low-cost, and commercially viable microwave sensor for early monitoring of high protein content in blood, which can act as a biomarker for highly fatal diseases like Cancer, Rheumatoid Arthritis, and human immunodeficiency virus (HIV). To achieve this objective, a highly sensitive, metamaterial-based sensor, incorporating novel hybrid circular-square shaped complementary split ring resonators (CSRR), at ISM band frequency range of 2.4 GHz is proposed. The effectiveness of the proposed sensor is validated on in-house developed blood-mimicking liquid samples placed in a microfluidic channel (polydimethylsiloxane). Minute changes in its dielectric properties, caused by an increase in its protein levels, are reflected in terms of the variation in the magnitude and frequency of the S21 response of the sensor. Specifically, the average transmission amplitude sensitivity (TAS) and the average frequency sensitivity (AFS) of the proposed sensor are 1.37 dB/(mg/ml) and 0.57 GHz/(mg/ml), respectively. Various machine learning classifiers were designed and analysed to categorise the simulation data of the proposed sensor into four different stages based on the protein concentration, among which the performance of the XGBoost-based classifier was found to be most effective. This work presents a proof-of-concept demonstration of the proposed microwave sensing methodology for potential real-world blood protein monitoring applications.